Jailbreaking LLMs Without Gradients or Priors: Effective and Transferable Attacks.

Zhakshylyk Nurlanov, Frank R. Schmidt, Florian Bernard

Browse the full ICPR paper archive.